AU665365B2 - Treatment of cell populations - Google Patents

Abstract

PCT No. PCT/GB92/01483 Sec. 371 Date Jul. 13, 1994 Sec. 102(e) Date Jul. 13, 1994 PCT Filed Aug. 10, 1992 PCT Pub. No. WO93/03151 PCT Pub. Date Feb. 18, 1993Disclosed is a method of treating a heterogeneous population of cells to link together copies of two or more nucleic acid sequences from at least some of the cells, the arrangement being such that copies of the DNA sequences from an individual cell are preferentially linked in the vicinity of the nucleic acid from which the copies are derived. Also disclosed are recombinant proteins expressed by the method of the invention and kits for performing said method.

Description

I- OPI DATE 02/03/93 AOJP DATE 13/05/93 APPLN. ID 24001/92 PCT NUMBER PCT/GB92/01483 I illl l ll llIAU9224 0 ll ll AU9224001 INItKNAIIUNAL APPLLAIIlUIN PUULI-HtU UNUtK 11-t VA I N 1 CLUUPbKA IUN I KEA IY (PCT) (51) International Patent Classification 5 (11) International Publication Number: WO 93/03151 C12N 15/13, 15/10 C12Q 1/68 A (43) International Publication Date: 18 February 1993 (18.07.93) (21) International Application Number: PCT/GB92/01483 (74) Agent: KEITH W. NASH CO.; Pearl Assurance House, 90-92 Regent Street, Cambridge CB2 1DP (GB).

(22) International Filing Date: 10 August 1992 (10.08.92) (81) Designated States: AT, AU, BB, BG, BR, CA, CH, CS, Priority data: DE, DK, ES, FI, GB, HU, JP, KP, KR, LK, LU, MG, 9117352.6 10 August 1991 (10.08.91) GB MN, MW, NL, NO, PL, RO, RU, SD, SE, US, Euro- 9212419.7 11 June 1992 (11.06.92) GB pean patent (AT, BE, CH, DE, DK, ES, FR, GB, GR, IE, IT, LU, MC, NL, SE), OAPI patent (BF, BJ, CF, CG, CI, CM, GA, GN, ML, MR, SN, TD, TG).

(71) Applicant (for all designated States except US): MEDICAL RESEARCH COUNCIL [GB/GB]; 20 Park Crescent, London WIN 4AL Published With international search report.

(72) Inventors; and Inventors/Applicants (for US only): EMBLETON, Michael, Jim [GB/GB]; 95 Derby Road, Bramcote, Nottingham NG9 3GW GOROCHOV, Guy [FR/GB]; 11 North Terrace, Midsummer Common, Cambridge (GB).

JONES, Peter, Thomas [GB/GB]; 37 High Street, Lode, Cambridge CB5 9EW WINTER, Gregory, Paul [GB/GB]; 64 Cavendish Avenue, Cambridge CB1 4UT

(GB).

6656(54 (54) Title: TREATMENT OF CELL POPULATIONS bac oligo V ri Ink oLigos st PCR(in cels) MoJk5For2 1st PCR( in cells)

BKNES

ols 77; iii~~ii~l~l 2nd PCR (after washing cells) :i (57) Abstract Disclosed is a method of treating a heterogeneous population of cells to link together copies of two or more nucleic acid sequences from at least some of the cells, the arrangement being such that copies of the DNA sequences from an individual cell are preferentially linked in the vicinity of the nucleic acid from which the copies are derived. Also disclosed are recombinant proteins expressed by the method of the invention and kits for performing said method.

WO 93/03151 PCT/GB92/01483 -1- Title: Treatment of Cell Populations Field of the Invention This invention concerns the treatment of cell populations and relates to methods of treating cell populations, e.g.

for analysing the linkage of genes within individual cells of a heterogeneous population, novel protein and nucleic acid compositions, and kits for performing said methods.

Background of the Invention This invention concerns techniques for the treatment of populations of cells, typically heterogeneous populations of cells, and has particular, but not exclusive, application to the treatment of lymphocytes such as those producing antibodies.

Antibodies are present in serum and bind to and help eliminate diverse pathogens such as toxins, bacteria and viruses. They consist of a Y-shaped protein built from two heavy chains and two light chains. Each chain has a modular construction: each light chain consists of two domains, and each heavy chain has at least four domains.

The antigen binding site is fashioned by one domain from the heavy chain (VH domain) and one domain from the light chain (VL domain). Indeed small antigen binding fragments can be prepared which consist of these two domains, either associated non-covalently, or via disulphide bonds 4 RA c I WO 93/03151 PCT/GB92/01483 -2or via a peptide linker. The antigen binding domains are more variable in amino acid sequence than the other domains of the antibody, and are therefore termed variable domains in contrast to the constant domains. The :constant domains of the antibody are responsible for jl triggering antibody effector mechanisms such as complement lysis and cell mediated killing.

Antibodies are made by B-lymphocytes in a process of gene rearrangement. During the development of these cells, the genes encoding the variable domains are assembled from J genetic elements. In the case of the VH domains there are three elements, the unrearranged VH gene, the D segment and the JH-segment. In the case of the VL domains there Sare two elements, the unrearranged VL (V lambda or V jI Kappa) gene and the JL (J lambda or J Kappa) segment.

Random combination of these gene segments and random combination of the rearranged VH and VL domains generates a large repertoire of diverse antibodies, capable of binding to diverse antigens. The potential diversity is large (primary repertoire in mouse is greater than 101), although the repertoire expressed at any time must be less than the total number of lymphocytes (in mouse Iss than 108.

In the first stage of the immune response, antigen is recognised by lymphocytes and they proliferate and secrete antibody which is usually of low affinity (less than 10 M 1 On further encounter with antigen (as in hyperimmunisation), the V-genes of these cells are subjected to mutation and lymphocytes displaying antibody 4 with enhanced affinity are selected. This process, in which high affinity antibodies are made in two stages from a primary repertoire and subsequently from a hyperimmune WO 93/03151 PCT/GB92/01483 3 repertoire is highly efficient. It has advantages over selecting antibodies in a single step, as this would require a much larger primary repertoire (the potential hyperimmune repertoire is perhaps greater than 1030).

Traditionally cell lines making high affinity antibodies have been prepared by fusing B-lymphocytes from a hyperimmunised animal with a myeloma cell line. This immortalises the B-lymphocytes and allows the cloning and screening of cells for production of monospecific antibodies binding to antigen. Such monoclonal antibodies have found a wide variety of uses in therapy and diagnostics. However the plasma cells (last stage of differentiation of B-lymphocytes) which are essentially factories for secretion of antibodies, do not fuse. Hence the repertoire of antibody specificities contributed by plasma cells is not .accessible by hybridoma technology.

This may explain why polyclonal antisera from animals can have higher average affinities than monoclonal antibodies derived from the same animal. There are other disadvantages of hybridoma technology, including the low efficiency of cell fusion, the time consuming steps 4 required to clone hybridomas with the desired antigen binding activities and sometimes the instability of hybridomas.

Recently other methods based on gene technology have been proposed for making antibodies from B-lymphocytes (see Milstein and Winter, Nature 349, 293-299 (1991) (reference 11) for review and references). The key step is the cloning of the genes encoding the VH and VL genes directly from B-lymphocytes (or hybridomas) into expression vectors.. The genes can be expressed as single variable domains, or Fv or Fab fragments or antibodies in either I Di' i, WO 93/03151 PCT/GB92/01483 4bacteria or mammalian cells. The key to these recombinant DNA based methods are the rapid cloning of the V-genes directly into expression vectors. This has been achieved by copying and amplifying the V-genes by the polymerase chain reaction (PCR) using "universal" primers which hybridise to the 5' and 3' ends of V-genes. The primers include restriction sites to permit the cloning of the genes directly into expression vectors.

These methods are highly suitable for making antibodies from clones of cells or indeed single lymphocytes.

However they do require separate amplification and cloning steps for each single lymphocyte or each individual clone, as the individual VH and VL elements have to be kept together. Alternatively it has been shown that antigen binding activities can be recovered by combining at random VH and VL genes taken from a large repertoire from an immunised animal. This relies on the fact that it may not be necessary to reproduce the precise VH and VL combinations of the B-lymphocyte to recover an antigenbinding activity. Indeed the VH and VL gene combinations derived from a large diverse hyperimmune repertoire will not generally correspond to the original combination of VH and VL genes present in B-lymphocytes. Suppose the VH and VL genes from only one thousand different lymphocytes were reshuffled. It would probably require the screening of over a million VH and VL gene combinations to have a reasonable chance of finding the original VH and VL gene combination from any one cell. This calculation assumes that the VH and VL genes of the chosen cell are unique: this will be particularly true for a hyperimmune response.

Thus VH and VL gene combinations selected from a large repertoire of V genes are likely to be artificial combinations. However it is possible that some artificial WO 93/03151 PCT/GB92/01 483 combinations with binding activities may be similar to the original combination, perhaps with a few amino acid substitutions. Nevertheless it is expected that the majority of such artificial antibodies will have lower affinities than the original combination (although a few may even have improved affinities). To improve their affinities it will generally be necessary to make further steps, for example by recombining one of the chains with a repertoire of diverse partners, or making mutations in the VH and VL genes and to selecting mutants with improved affinity (and thereby mimicking the process of affinity maturation in the hyperimmune response). It would therefore be desirable to clone the correct combinations of VH and VL elements together from heterogeneous and large populations of cells without the need for separate cloning steps for each cell, or without resorting to the entirely random pairing of VH and VL elements. It will also be necessary to clone the original combination of VH and VL genes if it is desired to study the natural antibody response to an antigen.

The present invention is based in part on the proposal that the combination of VH and VL genes present in each cell can be retained if they are copied and the copies of each are linked together preferentially. The linked copies could then be cloned directly into expression vectors. However depending on the efficiency of the linking and copying processes, it might be necessary to first make further copies of the linked genes before cloning.

One way of achieving linkage of the VH and VL genes is by in situ PCR, either of the genes themselves or of cDNA copies of their mRNA transcripts.

FRI,

7- 0 es WO 93/03151 PCT/GB92/01483 6 The principle of "in-cell" PCR is known and has been disclosed by, for example, Haase et al., 1990 (Proc. Nati.

Acad. Sci. USA 87, 4,971-4,975) (reference 15), and by Bagasra et al., 1992 (New Engl. J. Med. 326, 1,385-1,391) (reference 31). However, these papers are concerned with the use of in situ PCR to detect the presence of viral DNA sequences within infected cells. Haase et al. describe a system of PCR using overlapping primers to assemble already adjacent amplified sections of visna virus DNA so as to assemble a complete section of viral genome.

However, this is quite distinct from the present invention which involves the realisation that non-contiguous genomic sequences or cDNA may be linked (by PCR) preferentially within the same cell, thus preserving original combinations present within individual cells among a heterogeneous population.

Summary of the Invention In a first aspect the invention provides a method of treating a heterogeneous population of cells to link together copies of two or more non-contiguous DNA sequences from at least some of the cells, the arrangement being such that copies of the DNA sequences from an individual cell are preferentially linked in the vicinity of the nucleic acid from which the copies are derived.

The term "preferentially linked" is intended to indicate that those nucleic acid copies derived from a nucleic acid sequence within the same cell are more likely to become associated with each other than with a copy derived from a nucleic acid sequence from a different cell. Thus the original combination of genetic elements present in any one particular cell will tend to be conserved.

PCT/GB92/01483 WO 93/03151 20 WO 93/03151 PCT/GB92/G1 483 -7- This preferential linkage occurs in the "vicinity" of the nucelic acid sequence from which the copies are derived.

Use of this term is intended to indicate that the preferential linkage defined above occurs relatively locally, near the site of synthesis of the copied nucleic acids. The diameter of a mammalian cell is typically in the range 5-15um. Since the key factor is preferential linkage between nucleic acids copied from templates within the same cell, the volume or 'vicinity' within which linkage can occur preferentially varies according to the density of sites of nucleic acid synthesis. For reasons described below,, such foci or sites need not always be intact cells. The sites of synthesis should however generally be separated by at least one typical cell diameter (5-30um) because there may be some diffusion of the synthesised nucleic acid copies before linkage occurs.

Thus separation or "compartmentalisation" of the sites of synthesis will aid in maintaining the preferential nature of the linkage. These practical constraints require that linkage generally occurs within 30-60um of the site of synthesis, although as explained above, if the separation of the compartments within which synthesis takes place is increased, the effective volume available for preferential linkage is also increased. It There are two convenient methods of achieving this compartmentalisation, which allows for preferential i linkage. One is to perform the reactions within intact or substantially intact, cells, and the other is for the v reactions to be performed in discrete, physically separated localities.

The gene elements of cells could be encapsulated together and then the individual genes copied and the copies linked together within the capsule. The capsule must be robust S R enough to survive the copying process and retain the $1 SWO 93/03151 PCT/GB92/01483 S-8 individual gene copies sufficiently to allow linking of the gene copies from the same cell. The capsule can be permeable to reagents, such as oligonucleotide primers and polymerase, or impermeable in which case the reagents must be introduced into the capsule. The capsule could be the nuclear membrane, the plasma membrane or an artificial capsule constructed around the cell. The artificial capsule could be a separate artificial membrane or a different liquid phase, for example an organic solvent.

Thus the cell membrane might form the capsule if fixed with formaldehyde and made permeable to reagents.

An example of this is provided by in situ "in-cell" PCR.

The general technique of in-cell PCR is known and typically involves fixing stabilising with respect to temperature) and permeabilising the cells (so as to allow PCR reagents to enter the cell) and performing PCR in the conventional manner. There may be some diffusion of the PCR products into the external medium. However, in the method of the present invention, these products and excess PCR reagents may be removed by washing. i Alternatively the cells and reagents could be introduced i together into droplets and dispersed in an organic phase as an emulsion, in which case it is necessary to ensure that the majority of droplets are occupied on average by a single cell.

As a further possibility, if cells were dispersed away from each other, for example on a surface, such as glass j or plastic, or in an inert solid medium, such as agar or silica gel, or in a viscous medium such as glycerol or glucose solutions, there would be limited diffusion (or convection) of the gene copies from one cell to the next.

Therefore the gene element of one cell would tend to be linked to the gene element of the same cell, although

LI

CAc/AI yi0 nt I WO 93/03151 PCT/GB92/01483 9 there might be some linkage to the gene elements of other Scells.

Thus in a second aspect, the invention provides a method of treating a population of cells to link together copies of two or more non-contiguous genomic DNA or physically separated cDNA sequences from at least some of the cells, comprising treating the cells to stabilise them with respect to temperature and to permeabilise them with respect to reagents; adding primers for performing the polymerase chain reaction (PCR) such that the primers diffuse into the interior of the cells; subjecting the cells to suitable treatment so the DNA sequences of a particular cell are copied within that cell by PCR; an(" linking together the copies of DNA sequences within the vicinity of the cell from which they were copied.

The preferential linkage can be brought about by suitable techniques, e.g. by ligation using DNA ligase (with or without prior digestion with restriction enzymes).

However, most conveniently, the preferential linkage is brought about by PCR. This may be done by the use of sets of primers where an end of a primer which primes the synthesis of one nucleic acid sequence is complementary to an end of a primer which primes the synthesis of a second nucleic acid seqence.

Preferably the DNA sequence subjected to PCR is derived from cellular A'RNA. This is conveniently achieved by reverse transcription of the mRNA.

The use of fluorescently labelled PCR primers allows for the analysis of a population of cells for example by counting cells which express a particular gene or combination of genes.

I AVr O WO 93/03151 PCT/GB92/01483 Thus, in a further aspect the invention provides a method of treating a population of cells, comprising treating the cells to stabilise them with respect to temperature and to permeabilise them with respect to reagents; adding fluorescently labelled primers for performing PCR, such that the primers diffuse into the interior of the cells; subjecting the cells to suitable treatment so the nucleic acid sequences of a particular cell are copied by PCR; and monitoring the fluorescent labels.

A further aspect provides a method of treating a population of cells, comprising treating the cells to stabilise them with respect to temperature and to permeabilise them with respect to reagents; adding primers for performing PCR, such that the primers diffuse into the interior of the cells; sbjecting the cells to suitable treatment so DNA sequences of a particular cell are copied within that cell by PCR; adding fluorescently labelled probes; and monitoring the fluorescent labels.

Tyr .cally a cell population subjected to this tvthod of treatment is suitable for analysis by fluorescence activated cell sorting (FACS).

Fluorescence labelling of amplified DNA is particularly advantageous, as it can allow for the sorting of labelled cells by FACS) and the direct cloning of the amplified DNA from sorted cells.

Conveniently, two different primers labelled with different fluorescent labels are employed. This can allow for the analysis of two different sub-populations within a heterogeneous population in one experiment. Typically, i j WO 93/03151 PCT/GB92/01483 11 fluorescent labelling allows for counting of labelled cells, which is a preferred feature of this aspect of the invention.

Conveniently, cells may be stabilised with respect to heat by 'fixing' with formaldehyde. A convenient method of permeabilising cells is to treat them with a detergent, typically the non-ionic detergent Nonidet P40 It may be desirable that the methods defined above further comprise one or more rounds or further rounds of PCR.

Preferably any further rounds of PCR are performed using "nested" primers, which are complementary to sequences copied in the preliminary amplification.

A washing step may be usefully incorporated after the linkage step into any of the methods defined above, particularly those using in-cell PCR, to remove from the external surface or outer layers of cells unwanted material, e.g. linked copies from other cells, in order to reduce the background "noise". This is particularly useful if one or more further rounds of PCR are to be performed.

Any of the methods defind thus far are particularly suitable for application to the study of immunoglobulin VH and VL domains. Conveniently, therefore the nucleic acid sequences copied may be genomic DNA sequences (such as those encoding immunoglobulin VH and VL domains).

However, due to the higher effective concentration of mRNA (relative to genomic DNA) in a cell actively expressing a particular protein, it is preferable that mRNA may be reverse transcribed in situ into cDNA and the resulting L WO 93/03151 PCT/G B92/01483 12cDNA amplified by in cell PCR. Further, analysis of mRNA rather than genomic DNA more accurately reflects the activity of the cell.

The various method aspects of the present invention are also readily applicable to the study of the T cell receptor (TCR) V-genes in populations of cells.

Clearly, it may be beneficial to isolate the nucleic acid produced by the method aspects of the invention. Thus the products are typically inserted into suitable vectors. The inserted nucleic acid product is then generally sequenced or annealed to nucleic acid probes to confirm its identity. Particularly suitable are vectors which may express the nucleic acid sequence as a polypeptide. This is of special interest when isolating sequences encoding VH and/or VL domains, as it represents a method of producing mono-specific antibody binding fragments. Such cloning and expression vectors are well known to those skilled in the art. Particularly preferred is the technique whereby VH/VL combinations are expressed on the surface of a phage (McCafferty et al. such a technique representing a powerful selection method. 4 Expression on the surface of a phage is a particular embodiment of a generally preferred feature: expression of the cloned nucleic acid product as a fusion protein. As I an alternative to fusion with the coat protein of a phage, the cloned nucleic acid product could be expressed as a fusion protein with a peptide tag for detection purposes.

Thus the invention extends to proteins expressed by means of the method aspects of the invention. i WO 93/03151 PCT/GB92/01483 13 In a further aspect the invention provides a composition comprising copies of two or more nucelic acid sequences preferentially linked in the vicinity of the nucleic acid sequences from which the copies were derived.

In another aspect, the invention provides a kit for performing any of the method aspects of the invention, comprising at last a PCR primer for copying a DNA sequence of interest and intructions for use.

Optional components which may be included in the kit include further overlapping and/or nested PCR primers, labelled oligonucleotides which may be used as PCR primers or as probes, other PCR reagents (such as Taq polymerase, buffers) or reagents for fixing and permeabilising cells.

It will be apparent that the methods of the invention have a wide variety of potential applications. Normally, one would expect to employ such methods on heterogeneous populations of cells, but there is no reason why the method should not be used on a uniform population if it would reveal informat.on of interest. These applications include the construction of monoclonal antibodies (or fragments thereof) from populations of a B-lymphocytes, but it is by no means limited to this application. For example, it might be used for analysis of the combinations of chromosomes present in populations of somatic cell hybrids. Alternatively, instead of analysing diploid cell populations the method could be used to study haploid cell populations, such as sperm cells. Thus the method could be employed to inestigate, for example, linkage between genetic markers in human sperm, in a manner which improves upon that described by Li et al. (1988, Nature 345, 414- 417) (reference 39), as will be apparent to those skilled SUBSTITUTE SHEET tSP/EP WO 93/03151 PCT/GB92/01483 14 in the art.

The invention will be better understood by reference to the following illustrative examples and by reference to the accompanying drawings, in which: Figure 1 illustrates schematically certain steps of a method in accordance with the invention; Figures 2 to 5 show DNA sequences of oligonucleotides used as primers in Example 1; Figure 6 illustrates schematically assembly and nested amplification of 1g V-genes from hybridomas for cloning; Figure 7 illustrates schematically nested amplification of assembled 1g v-genes for fluorescent labelling; Figure 8 is a photograph of the results of gel electrophoresis of PCR products; Figure 9 illustrates intracellular localisation of amplified cDNA in cells using fluorescence labelling;and Figure 10 illustrates flow cytometry of K562 cells with bcr-abl amplified in accordance with a method of the invention.

Example 1 Linking of VH and VL genes from cells fixed with formaldehyde The VH and VL genes of a hybridoma cell line can be copied using the polymerase chain reaction and primers located at the 5' and 3' end of the VH and also of the Vk gene WO 93/03151 PCT/GB92/01483 15 (Figure The amplified VH and Vk DNA are also linked during the PCR reaction, in the order 5' VH-Vk 3' by virtue of a match between the primers at the 3' end of the VH gene, and the 5' end of the Vk gene (Stage To fish out the linked species from the unlinked species, a second set of 'nested' PCR primers is used (Stage 2).

The VH and Vk genes are linked preferentially within the same cell as follows. The cell is first fixed with formaldehyde to stabilise it to the temperature cycling, and treated with Nonidet P40 to permeabilise the cell clones analysed by probing and sequencing.

The nucleotide sequences of the mature heavy chain VH domains and the Vk light chain genes of the antiphenyloxazolone hybridomas NQ2/12.4 and NQ10/12.5, and the encoded protein sequences, are given in Figures 2-5. In Figures 2 and 4, the VH and Vk genes are linked by a nucleotide sequence encoding a polypeptide spacer providing a flexible polypeptide chain to link the two domains. Thus the complete nucleotide sequence in Figures 2 and 4 encodes a single chain Fv fragment. The spacer is incorporated within the PCR primers MoVhlnk3 and MoVklnk3, which are used with a VH back (NQ10/12.5 or NQ2/12.4) primer and a Vk forward primer (MoJkFOR2) to amplify and link the VH and Vk genes in a form suitable for expression. Figures 3 and 5 are similar, and show a shorter linker that contains a unique restriction site to allow the cloning of a longer polypeptide linker by introducing a synthetic oligonucleotide. To identify the linked genes, further cycles of PCR are used with a nested VH back (NQ10/12.5BKNES or NQ2/12.4BKNES) primer and a nested Vk forward (MokFORNES) primer.

SUBSTITUTE SHEET =Sf /P _1 I WO 93/03151 PCT/GB92/01483 16- In the example below, the sequence of the hybridoma genes was already known and primers specific to the hybridomas were designed. However in principle more general primers could be used, for example, to clone variable domains in general by PCR (reference Also the second round of PCR amplification in the example gives rise to a single chain Fv truncated at both N- and C-terminus. However, a different set of primers should overcome these problems.

For example, for the first stage PCR amplificationassembly, a family of Vk forward primers might be designed directly to the 3' side of the J-regions, either in the C kappa or C lambda constant domain for RNA templates, or in the region of DNA next to the splice junction for genomic DNA templates. Linking primers would be used that encoded a polypeptide sequence (as above), for example to join the VH and Vk as a single chain Fv fragment. The Vh back primer would incorporate a restriction site for cloning into expression vectors. For the second stage amplification of the linked VH and Vk genes, Vk forward primer(s) based within the Jk region itself (and incorporating restriction site for cloning into expression vectors) might be used in conjunction with the VH back primer.

Preparation of cell templates NQ2/12.4 or NQ10/12.5 hybridoma cells were grown in tissue culture flasks in Dulbecco's modified Eagle's medium containing 5% foetal bovine serum. Cultures in late logarithmic growth phase were harvested by rinsing adherent cells into the supernatant medium, and centrifuging the medium at 800 rpm for 5 minutes at room temperature. The supernatant was removed and the cells were resuspended in 5ml phosphate buffered saline (PBS, pH I IRqTITI ITF T= T qtSn (i WO 93/03151 PCT/GB92/01483 17- 7.2) at room temperature and again centrifuged at 300 rpm for 5 mins (1st wash). Two further washes in 5ml PBS were given under the same conditions. The cell pellets were then suspended in 10% formal saline (10% formalin 4% formaldehyde] and 0.15M NaCI in distilled water), using Iml per 10-50ml of original culture, and transferred to 2ml Eppendorf Safe-Lock tubes. They were incubated on ice for 1 hour, with frequent agitation on a vortex mixer.

The cells were then spun down at 13,000 rpm in a microfuge, and washed 3 times with 1ml ice-cold PBS, dispersing them by pipetting up and down several times at each wash. Following removal of the PBS after the 3rd wash, the pellets were resuspended in 1ml ice-cold Nonidet in water, and incubated on ice for a further 1 hour with frequent agitation. The cells were washed 4 more times in 1ml ice-cold PBS, with vigorous pipetting, and suspended in 0.2-0.5ml cold PBS containing 0.1M glycine.

A sample was examined microscopically in a haemocytometer, and if clumps were present they were dispersed into single cells by vigorous pipetting or repeated passage through a 26 gauge hypodermic needle. The ce]'s were then counted and adjusted to 107 per ml in PBS U.1M glycine, and aliquoted into 0.5 ml tubes (usually 0.05-0.1 ml per tube) and frozen in dry ice. The frozen aliquots were stored at degrees C.

Assembly and Amplification Heavy and light chain variable regions were assembled and amplified by a 2-stage Polymerase Chain Reaction. In the first stage, VH and Vk genes were linked together, and in the second stage the assembled product was amplified from the cell template. Reactions were carried out in Techne HI-TEMP 96 polycarbonate microplates using a Techne PHC-3 SUBSTITUTE SHEET r L SUBSTITUTE SHEET TS+ l6 WO 93/03151 PtGB2018 PC'r/GB92/01483 is programmable heating block, or in 0.5m1 Sarstedt tubes using a BioTherm Inc. BioOven.

Reaction conditions were as follows:- 1ST STAGE: Water Forward Vk primer Back Vh primer Forward Link primer Back Link primer dNTPs (5mM) x PCR buffer Cell template (in PBS/glycine) Tag polymerase (5 units/ul) 26 10.0 A drop of oil was added when Techrie plates were used, but none when tubes were used in the BioOven. Cycling times and temperatures were as follows: Techne PFIC-3 Block: B ioOve n: 94*C 30 secs 1 min 72*C 1 min 95 0 C 30 secs 30 secs 72 0 C 30 secs 30 cycles 30 cycles 1st stage Primers (10 pmol/ul) Vh Back NQ2/12.4 BACK GTG CAG CTG AAG GAG TCA GG 3' SUBSTITUTE SHEE, =F1/6 00=~ WO 93/03151 PCl'/GB92/01483 19 OR NQ1O/12.5 BACK AGC TGG TGG AGT CTG GGG G 3' Vk Forward MoJk5FOR2' S'CTT ACG TTT GAG CTC GAG CTT GG- 3' Forward Link MoVhlnk3 51CCA CTG CCG CCA CCA CCG CTA CCA CCA CCA CCT GCA GAG ACA GTG ACC AG 3' OR MoVhLNK4 TTT Ggc tag cTG GAG AGA GAG TGA (Nhe I) CCA GAG TCC CTT GGC CCC A 3' Back link MoVklnk3 1GCG GTG GTG GCG GCA GTG GCG GCG GCG GCT CTC AAA TTG TTC TCA CCC AGT CTC GAG C 3' OR MoVkLNK4 TGT CTC TGC gct agc cCA AAT TGT TCT GAG CCA GTC TCC AG 3' (The lnk3 primers were used as a pair, and likewise the LNK4 primers.) SUBSTIT ITP _QiL-r 01"v IQ& WO 93/03151 PCT/GB92/01483 CCA GAG TCC CTT GGC CCC A 3' jBack link MoVklnk3 5'GCG GTG GTG GCG GCA GTG GCG GCG GCG GCT CTC AAA TTG TTC TCA CCC AGT CTC CAG C 3' OR MoVkLNK4 TGT CTC TGC get age cCA AAT TGT TCT CAC CCA GTC TCC AG 3' (The lnk3 primers were used as a pair, and likewise the LNK4 primers.) After termination of the first stage, the cell templates were separated from the rest of the reagents and washed. In Techne plates, the supernatant PCR mix and overlying oil were carefully pipetted away, leaving the cells at the bottom of the well. The cells were then suspended in 0.2ml of PBS/0.1M glycine and spun down in a microfuge at 13,000 rpm. After resuspension in the same buffer they were again spun down for a 2Ad wash, then resuspended in 10ul PBS/glycine for use as the 2nd stage template. Tubes from the BioOven were spun at 13,000 rpm and the supernatant PCR mix removed, and the cells washed twice in 0.2ml PBS/glycine, before final resuspension in 10ul PBS/glycine for use as 2nd stage template. All 1st PCR supernatants were saved.

2ND STAGE Water 27.5 ul Forward nested primer 2.5 ul SUBSTITUTE SHEET FI e 4, SUSITT3EE WO 93/03151 PCT/GB92/01483 21 Back nested primer dNTPs (5 mM) x PCR buffer Washed stage 1 cell template Taq polymerase (5 units/ul) Cycling conditions were as in the included a final 10 minutes at 65 2nd stage Primers (10 pmol/ul) 2.5 ul 2.0 ul 5.0 ul 10.0 ul 0.5 ul 1st stage reaction, but degrees.

Forward nested primer AGC ACC GAA CGT GAG TGG 3' Back nested primer NQ2/12.4BKNES CCT CAC AGA GCC TGT CCA 3' NQ10/12, AGC CTG GAG GGT CCC GGA AAC 3' (used with respective cell template).

The products of both the 1st and 2nd stage PCRs were run in a 2% agarose gel, prepared in 0.5 x TBE buffer and containing 0.5 ug/ml ethidium bromide, using phiX174 HaeIII as a marker. The results indicated that in the first stage PCR, fragments of the size expected for single VH and Vk genes are seen, but after the second stage PCR fragments corresponding to the assembled fragments are seen, according to the schematic of Figure 1.

SUBSTITUTE SHEET. IE? :L i i i ~il~i WO 93/03151 PCT/GB92/01483 22- The putative sequences of the linked fragments are shown in Figures Example 2. Use of an emulsion for PCR amplification In this example, we show that it is possible to make a water in oil emulsion that is stable throughout the PCR cycle, and in which it is possible to amplify plasmid

DNA.

After trying a range of detergents, with HLB (Hydrophilic Liphophilic Balance) of 3.5 6 (from Triton-X15 to Brij52), and a range of oils (n-dodecane to heavy mineral oil (Sigma)), we found that Span-60 (Sigma) and light white oil (Sigma) gave an emulsion which survived at least rounds of PCR cycling. Droplet size was estimated as 25-50 um diameter sufficient to contain a cell.

Conditions suitable for PCR in emulsions are ,iven below.

The sequences of the primers are given in Example 1.

10 x PCR buffer [10 x PCR 100mm Tris-HCl pH 8.3/500mM, KC1/25mM MgCl2] 4ul 2.5 mM deoxynucleotide triphosphates (pH 2ul NQ10, 12.5 BACK primer 2ul MoJk5FOR2 primer (10pmoles /ul) 2ul MoVHLNK4 (Ipmole/ul) 2ul MoVkLNK4 (Ipmole/ul) Taq DNA Polymerase (Cetus lul plasmid containing NQ10/12.5 VH gene (3 ng pBluescriptIIKS fr(cn Stratagene) lul plasmid containing NQ10/12.5 vk gene (3 ng pBluescriptIIKS from Stratagene) (HPLC grade) to SUBSTITUTE SHEET

IP

WO 93/03151 PCT/GB92/01483 23 The mixture was incubated for 5 minutes at 640C.

100ul freshly made 2% Span 60 in Light White Oil was added at 64°C. The two phases were mixed by vortexing vigorously for 30 seconds then immediately transferred to the PCR block (Techne PHC-3) at 64'C and subjected to cycles of 94 deg. C 1 min, 62*C C 1 min, 72 deg. C 2 min, then held at 64 0 C. The emulsion was checked visually and appeared to be stable but had creamed. The excess oil phase was removed and remaining oil and Span 60 removed by one extraction at 64°C with diethyl ether, and then two further extracti. s at room temperature. The tubes were incubated at 64°C to evaporate the excess ether and 1 ul of the DNA was then removed for a second round of PCR with the nested primers as in Example 1. Analysis of the DNA on a 2% high-gelling temperature agarose gel demonstrated that assembly had occurred.

It is helpful to protect the Taq polymerase from denaturation that occurs in the microdroplets or in the formation of the droplets. Thus we have discovered the yields of assembled genes can be improved by using encapsulated enzyme, for example by using E. coli bacteria in which the enzyme has been cloned and is expressed (Engelke et al., Anal Biochem. 191, 396-400 (1990) (reference 40). By making alterations in the protocol above, within the skills of someone skilled in the art, it is to be expected that the VH and Vk genes could be amplified and assembled from populations of cells.

Examole 3 Reverse transcription of mRNA and amplification of cDNA.

SUBSTITUTE SHEET,.

/E

WO 93/03151 PCT/GB92/01483 24- Monoclonal antibodies have provided reagents for the identification of marker proteins (and lipids and carbohydrates) expressed within cells on the surface of cells or secreted from cells These reagents have had diverse uses, ranging from the localisation of cells in histological sections to the use of fluorescence activated cell sorting (FACS) of populations of cells However given the nucleotide sequence encoding a desired marker, we could in principle use the mRNA species as a marker for a cell. We have therefore explored a process in which the mRNA is reverse transcribed within the cell, the cFNA is amplified using the polymerase chain reaction and is detected in the cell using fluorescent PCR primers.

In principle such a process could also be used to detect and to clc.ie combinations cf mRNAs by linking the PCR amplified cDNAs within the same cell. In particular it could help in the cloning of Ig V-gene combinations from immunised or auto-immune sources using recombinant DNA techniques. At present, repertoires of antibody heavy and light chains are prepared from lymphocyte populations by PCR combined at random and expressed as soluble fragments in bacteria for screening with antigen or displayed on the surface of filamentous phage and selected by panning However the original combinations of heavy and light chains of the Blymphocytes are destroyed (11,12), and the artificial combinations can be dominated by promiscuous chains (10,13), leading to different affinities (10) and specificities In-cell PCR and linkage of heavy and light chains would allow the original combinations of the B-iymphobytes to be retained.

SUBSTITUTE SHEET "Mr.L' I PO WO 93/03151 PCT/GB92/01483 Earlier work had shown that lentivirus DNA can be amplified in situ using PCR, and the infected cells detected by autoradiography We have developed a similar approach to making and amplifying cDNAs within the same cell, as demonstrated with two hybridoma cell lines, B1-8 (16) and NQ10/12.5 and the K562 erythroleukaemia line The cells were fixed with formalin and permeabilised with the detergent NP40 to allow the access of nucleotides, primers and enzymes. For linking of the Ig heavy and light chain V-genes of the same cell during in-cell PCR, one primer was designed to prime at the 3' end of the VH gene, and another at the end of the VL gene, but with complementary tails to allow the self-assembly of the VH and VL genes.

Cell lines NQ10/12.5 murine hybridoma cells were grown in Dulbecco's modified Eagle's medium supplemented with 5% fetal bovine serum (FBS), and B1-8 hybridoma cells in RPMI 1640 medium supplemented with 5% FBS. Cells were used or passaged at late log phase, when almost all were viable. NQ10/12.5 secretes a kappa light chain and B1-8 a lambda light chain (16,17). The sequences of the VH and VL genes for both these hybridomas are known (16,17). The human myeloid leukemia cell line K562 was grown in RPMI with 5% FBS.

This line carries the Philadelphia chromosome, which is characterised by a bcr-abl fusion gene as a result of a translocation between the abl proto-oncogene on chromosome 9 and the bcr gene on chromosome 22 (19).

Cell fixation and permeabilisation

C)

T~V~o WO 93/03151 PCT/GB92/01483 26 Cells (107 to 108) were sedimented at 50 x g for 5 mins, washed 3 times in 5 ml PBS pH 7.2 at room temperature then suspended in 1 ml ice-cold 10% formaldehyde solution in 0.15M NaCI (formal saline). They were kept on ice for 1 hour, with occasional agitation on a vortex mixer. They were then spun at 13,000 rpm in a microcentrifuge for mins and washed 3 times in ice-cold PBS, with vigorous pipetting with a Pasteur pipette to resuspend and disperse any clumps visible by eye. The cells were next suspended in 0.5% Nonidet P40 (NP40, in water. After a further 1 hour incubation on ice with frequent agitation, they were again washed 3 times in ice-cold PBS with vigorous pipetting. A final wash was given in PBS containing 0.1 M glycine, and the cells were resuspended in 0.2 to 0.5 ml of the same buffer, using a 1 ml syringe and 26 gauge needle to disperse the clumps visible using a light microscope. The cells were counted, and for most experiments adjusted to a final concentration of 107 per 1. They were frozen in small aliquots in dry ice and stored at -7"C for up to a month.

Oligonucleotides The oligonucleotides are listed according to their use as primers or probes. The sequences were based on the sequence of the hybridoma V genes (16,17), and the bcr-abl gene and were synthesised using an Applied Biosystems 394 DNA synthesiser and used without purification.

cDNA synthesis MOLFOR CTT ACG TTT CAG CTC CAG CTT GG MOJH3FOR TAG GAC TCA CCT GCA GAG ACA GTG B1-8LFOR GCC TAG GAC AGT CAG TTT GGT TC B1-8VHLINK3 CCA CTG R A stJ IRTITI 1TF RHFFT P WO 93f 03151 WO 93/03151PC1/G B92/01 483 27 CCG CCA CCA CCG CTA CCA CCA CTG AGG AGA CTG TGA GAG TGG TGC first PCR MOLFOR, MOJH3FOR, B1-8LFOR, Bl-8VHLINK3, NQ2BK CAG GTG CAG CTG AAG GAG TCA GG NQ10BK TGC AGC TGG TGG AGT CTG GGG G B1-BBK CAG GTC CAA CTG CAG CAG CCT G BCR1A AGT TAC ACG TTC CTG ATC TC ABL2C TTA TCT CCA CTG GCC ACA AA 3' MOVHLINK3 CCA CTG CCG CCA CCA CCG CTA CCA CCA CCA CCT GCA GAG ACA GTG ACC AG MOVKLINK3 GCG GTG GTG GCG GCA GTG GCG GCG GCG GCT CTC AAA TTG TTC TCA CCC AGT CTC CAG C Bl-8VLLINK3 GCG GTG GTG GCG GCA GTG GCG GCG GCG GCT CTC AGG CTG TTG TGA CTC AGG AAT CTG C first PCR linker primers: Bl-8VLLINK3, Bl-8VHLINK3, MOVHLINK3, MOVLLINK3.

Second PCR (nested primers): AGC CTG GAG GGT CCC GGA AAC B1-8BKNES GAG CTT GTG AAG CCT GGG GCT T MOLFORNES CCC AGC ACC GAA CGT GAG TGG B1-8FORNES (51 CCA CCG AAC ACC CAA TGG TTG CT V186.2 AGA CAA ACC CTC CAG Second PCR (nested) tagged primers for fluorescence labelling: M13B1-BBKNES GAG GAA ACA GCT ATG ACC GAG CTT GTG AAG CCT GGG GCT M13B1-8FORNES GAG GAA ACA GCT ATG ACC CCA CCG AAC ACC CAA TGG TTG CT -21MOLFORNES TGT AAA ACG ACG GCC AGT CCC AGC ACC GAA CGT GAG TGG 21NQ10BKNES TGT AAA ACG ACG GCC AGT ACG CTG GAG GGT CCC GGA AAC -21BCRlB TGT AAA ACG ACG GCC AGT TCT GAC TAT GAG CGT GCA GA -21ABL2D TGT AAA ACG ACG GCC AGT AGT GCA ACG AAA AGG TTG GG Third PCR fluorescent-~primers: -21ROX or -21FAM TGT AAA ACG ACG GCC CAG M13ROX GAG GAA ACA GCT ATG AC were obtained commercially SUBSTITUTE SHEET 7W'IMP WO 93/03151 PCT/GB92/01483 28 from Applied Biosystems.

Probes: B1-8LPRB CTG TAC CAT AGA GCA CAG NQ10PRB GAG TTT CCG GGA CCC TCC AG NQ10KPRB TTG GAA CCA GTT CAT GTA C V186.2.

cDNA For in-cell cDNA synthesis of Ig V-genes of the hybridomas, 10 fixed cells were thawed, spun down and washed in 200 ul water, then resuspended in 20 ul water.

A 'first strand mix' was freshly prepared, comprising 5 ul x 1st strand buffer (1.4M KC1, 0.5M Tris-HCl pH8.1 at 42°C, 80mM MgCl 5 ul 0.1M dithiothreitol (DTT), 5 ul mM dNTPs, 25 pmol each forward primer, 80 units RNase inhibitor (RNasin, Promega) and water to a total volume of 28 ul. The cells were heated to 65°C for 3 mins then cooled on ice. The 'first strand mix' was added, followed by 40 units Super RT AMV reverse transcriptase (HT Biotechnology Ltd., Cambridge). The cells and reagents were mixed and incubated at 42 0 C for 1 hour, then the cells were spun down, washed in 200 ul PBS (pH 7.2) containing 0.1 M glycine (PBS/0.1 M glycine) and resuspended in 20 ul of the same buffer for use immediately in PCR. For K562 cells, cDNA synthesis was as above except using random primers and oligo (dT) (21).

For cDNA synthesis from hybridoma cell lysates, viable 6 cells (5 *x 106) were boiled for 5 mins in 100 ul water containing 0.1% diethyl pyrocarbonate and spun for 2 mins at 13,000 rpm in a microcentrifuge. A 'first strand mix' was prepared, comprising 10 ul 10 x 1st strand buffer, ul 5 mM dNTPs, 5 ul 100 mM DTT, and 25 pmol each forward SUBSTITUTE SHEET WO 93/03151 PCT/GB92/01483 29 I primer. It was added to a volume of 62-67 ul supernatant from the boiled cells, and the mixture heated at 65°C for 3 mins and left to cool at room temperature for 15 mins.

RNasin (160 units) and reverse transcriptase (100 units) were added to bring the total volume to 100 ul, followed by 1 hour incubation at 42°C. For PCR amplification, the reactions were set up as below, except with 5 ul of the soluble template per tube.

In-cell PCR from cDNA Reactions were set up in 50 ul volumes in 0.5 ml Sarstedt tubes with 10 ul fixed template cells in PBS/0.1 M glycine buffer, 25 pmol back primer, 25 pmol forward primer, 200 uM dNTPs, 5 ul lOx Taq polymerase buffer (Promega) and units of Taq polymerase. The tubes were subjected to up to 40 cycles of PCR with denaturation at 95°C and extension at 72°C for 1 min each. Annealing was for 1 min at temperatures ranging from 60°C to 72C. For situations 4 requiring cell recovery, a thermal cycling oven ("BioOven", BioTherm Corp.) was used, which did not require the use of a mineral oil overlay. Where cell recovery was not required, cycling was performed on a cycling heat block (Techne PHC-3) with an overlay of 50 ul mineral oil.

In-cell PCR and assembly from cDNA The cDNAs were amplified by PCR and linked in the same reaction by using primers with complementary tails (22) (Fig.

L

Figure 6 illustrates assembly and nested amplification of Ig V-genes from hybridomas for cloning, cDNA synthesis and SUBSTITUTE SHEET /EP WO 93/03151 PCT/GB92/01483 30 assembly of VH and VL gene from hybridomas NQ10/12.5 and Bl.8. Sequences of primers (small arrows) indicated are as described previously. In each step the template strand is marked with a thick line, and the newly synthesised (first) strand with a thin line; in some cases, two steps are combined in the figure, and the second synthesised strand (using the first as template) is then marked with a dashed line. Sequence tags for assembly are marked as blocks. cDNA synthesis first PCR, further rounds of first PCR giving rise to VH and VL with "tagged" and complementary ends, assembly during first PCR of VH and VL, second "nested" PCR, assembled and amplified VH and VL genes. Note that for the cDNA synthesis from B1-8 cells, the tagged primer B1-8VHLINK3 was used, but the tag is not marked on the figure.

Reactions were set up in 50 ul volumes in tubes as follows: pmol VH back primer, 25 pmol VL forward primer, 10 pmol VH forward primer with linker sequence, 10 pmol VL back primer with linker sequence, 200 uM dNTPs, 5 ul 10 x Taq polymerase buffer, 2.5 units Taq polymerase, and 10 ul 55 fixed cells in PBS/0.1 M glycine buffer. Generally 105 (but sometimes up to 5 x 10 cells per tube were used, and the tubes were given 30 cycles of 95*C for 30 secs, for 30 sees and 72"C for 30 secs. The cells were spun down at 13,000 rpm, washed twice in 200 ul PBS/0.1 M glycine, and resuspended in 10 ul PBS/glycine. To amplify the assembled products, a second PCR was set up with the washed cells, nested primers (23) using 25 pmol nested VH back primer and 25 pmol nested VL forward primer, 200 uM dNTPs, 5 ul 10 x PCR buffer and 2.5 units Taq polymerase.

The cells were subjected to 30 more cycles of PCR, and DNA for cloning was isolated from the supernatant.

1 SUBSTITUTE SHEET "ZS /eP WO 93/G3151 PCT/GB92/01483 31 SIn experiments involving a mixture of two cell lines, all four PCR primers for each cell line (a total of eight primers) were included in both first and second "nested" PCRs using a 100 ul reaction volume. cDNA was synthesised from a 1:1 mix of Bl-8 and NQ10/12.5 (NQ10) fixed cells using the primers MOLFOR, MOJH3FOR, Bl-8LFOR and Bl- 8VHLINK3, and the cells washed and resuspended in PBS/0.1 M glycine for PCR assembly. The first PCR was carried out using the VL forward primers MOLFOR and Bl-8LFOR and the VH back primers NQ10BK and Bl-8BK. To link the two sets of genes, the PCR linker primers MOVHLINK3, MOVKLINK3, Bl- 8VLLINK3 and B1-8VHLINK3 were also included. The 3' ends of the PCR linker primers are complementary to the erds of the DNA template, but incorporate a 5' "tag" to allow assembly of the VH and VL genes (Fig. The tags were identical for both NQ10 and B1-8 linker primers, so that all possible combinations of VH and VL could be formed.

The second PCR used the nested primers MOLFORNES, Bl-8BKNES and Bl-8FORNES.

Fluorescence detection of intracellular PCR products cDNA in the B1-8 and NQ10 cells was subjected to PCR to amplify VH genes, using the primers indicated in Table 1.

They were then washed twice and the genes further amplified with nested PCR primers tagged with sequences complementary to those fluorescence tagged primers 21ROX, -21FAM, M13ROX used for DNA sequencing) (Table 1).

After washing, a third PCR was used with 3pmol of M13 or 21 primers labelled with 6 carboxy-X-rhodamine (ROX) (Applied Biosystems). Conditions for each PCR were 940 C for 40 secs, 58'C for 40 secs, and 72°C for 40 secs.

After the final stage the cells were washed three times in SUBSTITUTF SHFFT =Sl IEP

RWMWI

WO 93/03151 PCT/GB92/01483 32 TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 7.4) and suspended in the same buffer. A similar protocol was used to amplify assembled VH-VL DNA (Table the cDNA of K562 cells was subjected to amplification using PCR primers specific for the bcr-abl fusion gene (Table 1).

For fluorescence microscopy of K562, ROX labelled primers were used in the third PCR, and for flow cytometry both ROX and 5'-carboxyfluorescein (FAM) labelled primers (Applied Biosystems) were used in separate preparations (Fig. 7).

Figure 7 illustrates nested amplification of assembled Ig V-genes for fluorescent labelling. In this Figure, (1) shows second "nested" PCR with tagged primers to introduce primer sites for "universal" fluorescent primers and (2) shows third PCR with fluorescent primers. The same fluorescent primer can prime at both the 5' and 3' ends of the tagged sequences.

The cells were examined and photographed on a MRC 500 3 scanning laser confocal microscope (BioRad). Between 10 4 and 10 cells in 10 ul TE were spread on a glass slide and covered with a coverslip, sealed at the edges with nail varnish. In addition, cytofluorographic analysis was performed on a FACScan flow cytometer (Becton Dickinson).

Fluorescence data were displayed as logarithmic overlay histograms using Consort 30 software (Becton Dickinson).

Cell debris was gated out and 105 cells were analysed per sample.

Cloning of assembled genes DNA from the supernatants of cells after assembly and amplification was purified from 1.5% agarose gels in 0.5 x 0_ _SUBSTITUTE SHEET TSA ISP WO 93/03151 PCT/GB92/01483 33 Tris-borate EDTA (TBE) buffer (pH 8.0) with ethidium bromide (0.5 ug/ml) incorporated into the gels. The assembled DNA bands of around 600 bp were eluted with Geneclean II (BIO 101 Inc.), with the addition of Geneclean 'TBE modifier'. The purified DNA was ligated into a 'T vector' derived from Bluescript KS+ (Stratagene) and the vector transfected into E. coli CMK603 by electroporation. The cells were grown overnight on TYE plates (25) with 100 ug/ml ampicillin, 5-bromo-4-chloroindolyl-B-D-galactoside (26) and isopropyl-B-D-l-thiogalacto-pyranoside White colonies were replica plated in order to prepare colony lifts for hybridisation, or for use as templates for PCR screening.

For PCR screening of colonies the reactions were set up in 20 ul volumes in individual wells of Hi-temp plates (Techne), containing 8 pmol of forward and back primers, 200 uM dNTPs, 2 ul 10 x Taq buffer and 0.4 units Taq polymerase per well. For example to identify clones from the in-cell PCR of mixed hybridomas NQ10 and Bl-8, each clone was screened with the primers NQO1BKNES and MOLFORNES, BI-8BKNES and Bl-8FORNES, NQ10BKNES and Bl- 8FORNES, and Bl-8BKNES and MOLFORNES. Each well was inoculated with bacteria from a single replica plated colony by means of a toothpick, and overlaid with 50 ul of mineral oil. The plates were subjected to 30 cycles of for 30 secs, 65'C for 1 min and 72*C for 1 min on a Techne PHC-3 heat block designed for multi-well plates.

The PCR products were run on 1.5% agarose gels (10 ul per lane).

For colony hybridisation, bacterial colony lifts were made using Hybond-N nylon membranes (Millipore). The bacteria were lysed with sodium dodecyl sulphate and the DNA Rl IRTITI IT= QWCL 7SA IPO WO 93/03151 PCT/GB92/01483 denatured and fixed to the membrane by microwaving (29), followed by UV cross-linking. Oligonucleotide probes for heavy (V186.2) and light (Bl-SLPRB) chains of the Bl-8 hybridoma, and heavy (NQ10PRB) and light (NQ10KPRB) chains 32 of the NQ10/12.5 hybridoma, were labelled with gamma p- ATP using polynucleotide kinase (New England BioLabs) and hybridised to the nylon filters for 16 to 60 hours. The filters were washed with solutions containing tetramethylammonium chloride (21) and exposed on Kodak XAR film.

In-cell PCR from mRNA (results) In earlier work, cells infected with lentivirus had been fixed with paraformaldehyde and stored in ethanol We fixed the hybridoma and leukaemia cells with formal saline, permeabilised them with NP40, and stored the cells frozen in PBS/0.1 M glycine. We found that with these cells, our method resulted in high yields of amplified DNA as detected in the cell supernatant (see below). The fixed and permeabilised cells remained intact during temperature cycling, and in PCRs from genomic DNA with primers specific for the VH and Vk regions of the hybridoma lines NQ2/12.4 and NQ10/12.5 (NQ10), amplified products of around 300 bp were routinely obtained in the cell supernatants. Permeabilisation was a necessary step (Fig. but formaldehyde concentrations ranging from 4% to 10% were equally effective for fixation. However, fixed and permeabilised cells which had undergone cDNA synthesis provided a much better template for PCR amplification (as shown by the examples in Fig. which shows the gel electrophoresis of PCR products. In this Figure, supernatants from cDNA amplifications were run on a 1.5% agarose gel in TBE buffer. These are from left to R1 IR.qTr[T rrc L- Lnm- 0 WO 93/03151 PCT/GB92/01483 35 right; lanes 1 and 11, phi X174 HaeIII marker; lane 2, PCR of Bl-8 VH from fixed cells (formalin only); lane 3, PCR of Bl-8 VH from fixed and permeabilised cells (formalin and NP40); lane 4, PCR assembly of B1-8 VH and V lambda from fixed and permeabilised cells; lane 5, as lane 4 except using nested primers in second PCR; lane 6, as lane except with fixed cells; lane 7, blank; lanes 8, 9 and as lanes 3, 4 and 5 but with NQ10/12.5 hybridoma.

Consistently higher yields of amplified DNA were obtained when cells were added to the reaction in their storage buffer (PBS/0.1M glycine) rather than water. In cells subjected to two-stage PCR assembly in which 10 uCi pmol) 35S-dATP was added to label the first PCR, of the radiolabelled product was found in the supernatant after the second PCR and 30% was retained within the cells.

The amplified DNA within the cells could be visualised directly by confocal microscopy (or conventional fluorescence microscopy) by incorporation of a fluorescent PCR primer in a third PCR step Confocal microscopy was used to detect the cDNA amplified with fluorescent PCR primers (see Table 1 for details of primers and controls). The results are shown in Figure 9 in which shows PCR amplification of VH gene in NQ10/12.5 cells, shows amplification of VH gene in Bl- 8 cells, shows PCR assembly and nested amplification of VH and Vk genes in NQ10/12.5 cells, shows PCR amplification of VH gene in NQ10/12.5 cells, and tagging using -21NQ10 primers in the second PCR, and M13ROX in the third PCR, shows PCR assembly and nested amplification of VH and V lambda in B1-8 cells, and shows PCR ,A4/ SUBSTITUTE SHEET "S /EP WO 93/03151 PCT/GB92/01483 35 amplified bcr-abl in K562 cells.

Flow cytometry of K562 cells with amplified bcr-abl was also carried out, and the results are shown in Figure for PCR amplification and labelling of cells with specific or non-specific (M13) ROX or FAM labelled primers, as in Table 1. In the Figure, shows excitation detected in the red channel; -21ROX (solid line) and M13ROX (dotted line) and shows excitation detected in the green channel; FAM labelled primers -21FAM (solid line) and M13FAM (dotted line).

The correct combinations of primers for NQ10/12.5 produced strong fluorescence in NQ10/12.5 (and not in Bl-8 cells), and vice versa for B1-8 (Table 1, Fig. The fluorescence was brighter for PCR assembled DNA than for PCR amplified VH DNA, perhaps due to the use of two fluorescent PCR primers or the greater retention of the longer DNA species within the cell Although there were differences in fluorescence intensity among different cells, most of the cells amplified with the correct primers were fluorescent. The bcL-abl fusion gene in K562 cells subject to in-cell PCR could be detected by fluorescence microscopy (Fig. and also by fl,.w cytometry (FACS) (Fig. 10). The fluorescence was located in the cytoplasm, and FACS analysis confirmed that the majority of the cells were fluorescent. Amplification of the bcr-abl gene using the chosen primers confirmed the synthesis of cDNA in-cell prior to PCR, since the amplified portion (318 bp) contained three introns (21).

In-cell PCR assembly Bl-8 and NQ10 cells were fixed, permeabilised and mixed in Tr S SUBSTITUTE SHEET ,,SP /EP WO 93/03151 PCT/GB92/01483 37 equal ratios. The mixtures were then treated to cDNA synthesis and PCR amplification and assembly (Fig. and the linked product isolated from the supernatant and cloned. The clones were then screened for the four different combinations of heavy and light chains (as described elsewhere). The results of three independent experiments are shown in Table 2 for 1:1 mixtures. In all cases the in-cell PCRs resulted in linkage of NQl0 VH with V Kappa, or B1-8 VH with V lambda, corresponding to the combinations of the hybridomas; no "crossover" were seen in the 104 clones evaluated. By contrast, .JRs performed with soluble cDNA led in addition to cross-over combinat.ons of NQ10 VH with V lambda and BI-8 VH with V Kappa.

These findings were confirmed for colonies from the first experiment in Table 2, by hybridising blots of replica 32 plated bacterial colonies with P labelled oligonucleotide probes for internal sequences of the four V-genes (V186.2, B1-8LPRB, NQ10PRB and NQ10KPRB). Again in-cell PCRs gave rise to clones with the combinations of the hybridomas, while clones from the soluble cDNA resulted in additional cross-over combinations (Table 2).

A more stringent test of VH and VL linkage was performed by PCR of cell and soluble cDNA templates from a mix containing 10% Bl-8 cells and 90% NQ10/12.5 cells (Table In this case, 450 clones of each PCR assembly were 32 initially probed with P kinase labelled V186.2 oligonucleotide probe to identify those clones expressing the Bl-8 VH chain. The positive clones were then screened on agarose gels to detect either V Kappa and V lambda genes using PCR with the primers MOVKLINK3 and MOLFORNES, and B1-SVLLINK3 and Bl-8FORNES respectively. From the in- |<RV|qT!TlT= CUCCt ea ia WO 93/03151 PCT/GB92/01483 30 cell PCR, all the Bl-8 VH clones (27) were all linked to the V lambda chain. From the assembly from soluble cDNAs, 2 of the 34 B1-8 VH positive clones were linked to the V lambda chain, and 31 to the V Kappa chain.

Example 4 Linking of Human Immunoglobulin Heavy and Light Chain Variable Domain Genes from a Mixed Lymphocyte Population We describe a further example linking the human immunoglobulin (Ig) variable region genes (VH and VL) within single cells in a lymphocyte population, such that the assembled products can be cloned into a vector and expressed as soluble antibody fragments or displayed on the surface of phage. In essence, it comprises synthesis of cDNA using forward primers annealing to the Ck gene and the JH segment, followed by assembly with linker primers, VH back primers based on the VH3 leader sequence, and a forward Ck primer nested with respect to the primer used for cDNA (Figure 11). The assembled product within the cells is then amplified with nested primers annealing to the 5' end of the VH gene and the 3' end of the Jk segment.

Cells Heparinised human blood was layered onto Lymphoprep lynphocyte separation medium (Nycomed) in a centrifuge tube, using equal volumes of undiluted blood and Lymphoprep. Tubes were centrifuged at 4,000 x g for minutes, after which mononuclear cells (which include lymphocytes) were visible as a band separating the blood plasma (upper layer) and the Lymphoprep (lower layer).

S T1 SUBRTrTI IeT c nu,- es M U WO 93/03151 PCT/GB92/01483 39 The cells were removed and washed 3 times in phosphate buffered saline (PBS, pH and incubated in formaldehyde in 0.15M NaCI solution on ice for 1 hour.

They were again washed in PBS 3 times, followed by incubation on ice in 0.5% Nonidet P40 (BDH) in water.

After a further 3 washes in PBS the cells were suspended in PBF containing 0.1M glycine and counted. They were stored frozen at Oligonucleotide primers The following primers were used in cDNA first strand synthesis and the PCRs. Their annealing positions relative to the VH and VL genes are depicted in Figure 11.

VH3LDR1 5' GCT (A/C)TT TTA A(G/A)A CGT GTC CAG TGT 3' VH3LDR2 VH3LDR3 VH3a VH3aSfi JH6AFOR CkEXTF JklFOR Jk2FOR Jk3FOR Jk4FOR JklFORNot GGT ATT TTA CAA GGT GTC GCT ATA TTA (A/G)(A/G)A GAG CTG CAG CTG CTG GAG GTC CTC GCA ACT GCG GCC GAG CTG CAG CTG CTG GAG GAA CCC TGG TCA CCG TCT GGA CCA CGG TCA CCG TCT GGA AGA TGA AGA CAG ATG CAG ATT TCA ACT GCT CAT ACG TTT GAT TTC CAC CTT ACG TTT GAT CTC CAG CTT ACG TTT GAT ATC CAC TTT ACG TTT GAT CTC CAC CTT ACG TTT AAT CTC CAG TCG GAG TCA TTC TCG ACT TGC TTC CAC CTT GGT CCC 3' CAG TGT 3'

GGT

TCT

CAG

TCT

CCT

CCT

GTG

CAG

GGT

GGT

GGT

GGT

TGT

GGC

GCC CAG TGT 3' GG 3' CCG GCC ATG GCC GG 3' CAG GTG G 3' CAG GTG C 3' CAG 3' ATG G 3' CCC 3' CCC 3' CCC 3' CCC 3' CCC 3' CGC ACG TTT GAT Jk2FORNot 5' GAG TCA TTC TCG ACT TGC GGC CGC ACG TTT GAT SUBSTITUTE SHEET %E

P

WO 93/03151 PCT/GB92/01483 40 Jk3FORNot 5' Jk4FORNot 5' 5' CTC CAG GAG TCA ATC CAC GAG TCA CTC CAC GAG TCA CTC CAG CTT GGT CCC 3' TTC TCG ACT TGC GGC CGC ACG TTT GAT TTT GGT CCC 3' TTC TCG ACT TGC GGC CGC ACG TTT GAT CTT GGT CCC 3' TTC TCG ACT TGC GGC CGC ACG TTT AAT TCG TGT CCC 3' JHLINK 1,2 5' CCA GAG CCA C GAG GAG ACG G' JH3LINK 5' CCA GAG CCA C GAA GAG ACG G' JH6LINK 5' CCA GAG CCA C GAG GAG ACG G' VkLINK 5' CAG GCG GAG G' AAA TTG TGT T VkLINK1 5' CAG GCG GAG G' ACA TCC AGA Ti cDNA first strand synthesis CT CCG TG ACC CT CCG TG ACC CT CCG rG ACC TG GCT GA CGC TG GCT GA CCC CCT GAA CCG CCT CCA CCT AGG GT(C/T) CC 3' CCT GAA CCG CCT CCA CCT ATT GTC CC 3' CCT GAA CCG CCT CCA CCT GTG GTC CC 3' CTG GAG GTG GCG GAT CGG AGT CTC C 3' CTG GAG GTG GCG GAT CGG AGT CTC C 3' Aliquots of 5 x 105 fixed and permeabilised cells were thawed and washed once in water, and resuspended in water. The following "first strand" mix was prepared:- 10 x 1st strand buffer (1.4M KC1, 0.5M Tris-HCl pH 8.1 at 42°C, 80mM MgCl 5ul 0.1M dithiothreitol, 5ul dNTPs, 25 pmol each primers JH4-5A, JH6AFOR and CkEXTF, units of RNase inhibitor (RNasin, Promega) and water to 28ul. The cells were heated to 65C for 3 minutes and cooled on ice. They were then added to the 1st strand mix together with 40 units of Super RT AMV reverse transcriptase, and the whole mixture incubated at 42 0 C for 1 hour. The cells were then spun down, washed once in PBS/0.1M glycine and resuspended in the same buffer for SUBSTITUTE SHEET %SA /EP S4 -zIKUA or -zIFAM TGT AAA ACG ACG GCC CAG M13ROX S* CAG GAA ACA GCT ATG AC were obtained commercially A SUBSTITUTE SHEET T P WO 93/03151 PCT/GB92/01483 41

PCR.

PCR

Reactions were set up in 50ul volumes in 0.5 ml Sarstedt tubes, using 105 cDNA template cells in 10ul together with a mix containing 25 pmol VH back primer mix (VH3LDR1 VH3LDR2 VH3LDR3), 25pmol Ck5'FOR, 10 pmol linker primer mix (JHLINK 1,2 JH3LINK JH6LINK VkLINK VkLINK1), 200 uM dNTPs, 5ul 10x Taq polymerase buffer (Promega) and unit Taq polymerase. The tubes were subjected to cycles of PCR with denaturation at 95°C for 30 seconds, annealing for 58 or 65°C for 30 seconds, and extension at 72°C for 30 seconds, in a thermal cycling oven (BioTherm Corp. "BioOven") without an oil overlay. This was referred to as the "1st PCR".

The cells were spun down and washed twice in PBS/0.1M glycine, and suspended in 10ul of this buffer for a 2nd PCR together with the following mix:- 200 uM dNTPs, 5ul 10x Taq buffer, 2.5 units Taq polymerase, and either of primer combinations 25 pmol VH3A and 25 pmol of a mix of JklFOR Jk2FOR Jk3FOR Jk4FOR Jk5FOR, or 25 pmol VH3ASfi and 25 pmol of a mix of JklFORNot Jk2FORNot Jk3FORNot Jk4FORNot Water was added to bring the cells and reagents to 50ul, and PCR cycling was carried out as for Sthe first stage. When primer combination was used in the 2nd PCR, the cells were recovered and washed, and treated to a 3rd PCR with primer combination to append restriction sites for cloning.

The PCR products were analysed by running on 1.5% agarose gels in 0.5 x Tris-borate-EDTA buffer containing SUBSTITUTE SHEET TV /E WO 93/03151 PCT/GB92/01483 42 ethidium bromide and viewing under ultra-violet light.

RESULTS

Assembly using a two-PCR protocol with primers incorporating restriction sites in the 2nd PCR (primer combination or a three-PCR protocol with primer combination in the 2nd PCR and in the 3rd, resulted in a band of DNA of the size expected for an assembled VH3 Vk antibody fragment (about 700 bp).

Since this fragment would include VH3 Sfi and Vk Not restriction sites, it could be digested with the appropriate enzymes and cloned into the vector pHEN (references 38, 41). This would allow expression as cloned fusion products displayed on the surface of phage if introduced into a suppressor strain of E. coJ such as TG1, or as soluble single chain Fv fragments with a C terminal peptide tag following infection of a nonsuppressor strain sch as HB2151 Although this example made use of primers designed for k light chains and the VH3 family the process is equally applicable to other human heavy and light chain families, and primers could be designed accordingly (38).

Discussion of Examples Earlier work had shown that the DNA of lentivirus infected cells could be PCR amplified in infected cells and detected by auto'radiography and more recently HIV provirus has been detected by PCR amplification and peroxidase staining in cells attached to glass slides In-cell PCR allows the genes of individual cells to be amplified, but without the need to place each cell in a .separate tube It partitions the amplification of a SUBSTITUTE SHEET

I

/EP

WO 93/03151 PCT/GB92/01483 43 cell population, allowing many independent reactions in the same tube. We have extended the use of in-cell PCR by reverse transcription of mRNA to cDNA within single cells of a suspension (Example The cDNA is retained within the cell and allows PCR amplification. Furthermore the amplified DNA could be fluorescence labelled within the cell using fluorescent PCR primers, and should allow the analysis of large and diverse populations of cells by fluorescence. Here we used three PCR amplifications in order to tag the amplified DNA, and then fluorescence lable and tagged DNA with "universal" fluorescent PCR primers. However, it should also be possible to fluorescence lable the cDNAs within the cell with a single PCR amplification using a set of fluorescent PCR primers specifically made for the gene of interest.

In-cell PCR from cDNA appears to be efficient, as most of the c.lls of the hybridomas and K562 myeloid leukaemia could be fluorescence labelled, as determined by confocal microscopy (Fig. 9) or FACS (Fig. 10). A potential advantage of fluorescence labelling over use of I radiolabelled probes is that rare positive cells could be isolated by FACS, and their genes cloned directly or after further rounds of PCR.

We were able to assemble the PCR amplified DNA from two mRNAs of the same cell and to amplify the assembled DNA using nested primers, allowing tlhe combinations of linked genes to be identified by fluorescence (Table and to be cloned. For mixtures of two hybridomas, the original combinations of the hybridoma V-genes were retained even with one hybridoma in ten told excess. The extent of linkage between two markers of one cell is likely to depend inter alia on whether both mRNAs are amplified SUBSTITUTE SHEET TS= E SUBSTITUTE SHEET /eP WO 93/03151 PCT/GB92/01483 44efficiently and to high copy number, and on whether there is competition from amplified DNA leaking from other cells to take part in the assembly process. After the first PCR, we therefore washed the cells to remove DNA from their surface or outer layers before using the second "nested" PCR to amplify those V-genes assembled within individual cells. The strong linkage we see between the V-genes within the same cell may allow even rare combinations of genes to be rescued from a large population of cells.

In-cell PCR and assembly has many potential uses for gene linkage analysis, but a major application is likely to be the cloning of gene combinations that are polymorphic within a population of cells, such as the rearranged genes for Ig or TCR V regions. The linked genes could be cloned for probing, sequencing or expression. Thus in example 4, the way we have assembled the V-genes allows the cloning of the linked product directly into expression vectors.

Such a process would provide alternatives to B and T-cell hybridomas, allowing the original V-gene combinations to be cloned directly into vectors for expression in mammalian cells or bacteria (33,34,35,36,37) or displayed on the surface of filamentous bacteriophage By antigen panning of filamentous bacteriophage displaying antibody fragments, even very rare clones could be isolated and this should facilitate analysis of the V-gene combinations, smatic mutation, affinities and specificities of natural, immune and auto-immune B-cells.

I

M

OZ

Y:

W

r^j^ WO 93/0315 1 45 Table 1. In cell POR with fluorescence labelled pritners PCT/G B92/01 483 cells Primers PCR1 Primers PCR2 Primers PCR3 Fluorescent Cells NQIO NQ10BK -21NQIOBKNES -21 ROX j (VH gene) MOJH3FOR MOMHFOR M13 ROX NQ1 0 BI-8EK M13Bl-8BKNF-S -21 ROX- (VH gene) B 1-8VHLNqK3 V 186.2 M13 ROX f B 1-8 NQIOBK .21NQIOBKNF-S -21 ROX (VH gene) MOMhIFOR MOJH3FOR M13 ROX B 1-8 B1-8BK M13B1-8BKNES -21 ROX (VH gene) BEI .VHLUI4K3 V186.2 M13 ROX -21 ROX NQ1 0 MOVHLD;UX *21NQ10E1GNJES (VHVKgens)MOVKLNK3 -21MOLFORNP-S MOLFOR M 13 ROX B1-8VHBK 2RO BI8B1-8VHUNK M13BI-8BKNES-2RX EVl- gees B 1-8VLLINK3 M M13B I-8FORNES (VH/B gee I B-8LFOR M M13 ROX K562 ABL2C -21BCRB -21ROX (bcr-abt) BCR1A -21ABL2D M13 ROX Y,56 2 ABL2C -21BCRE -21 PAM Cbcr-abl) ECRIA *21ABL2D M M13 PAM- WO 93/03151PC/B2048 PCT'/GB92/01483 46 Table 2. Screening of linked VH and VL gene combinations from 1:1 mixtures of I-S and NQ10/12.5 hybridorna cells.

PCR

templ~te combination Expt. 1 Expt. 1 (probe) Expt. 2 EXPt. 3 Fixed cells NQ10 VH BI1-8 yR B1-8 V% NQ1O VH B 1-8 VX B 1-8 VH 20/47 27/47 0/47 0/47 19/32 13/32 0/32 0/32 11/36 25/3 6 0/36 0/36 11/2 1 10/21 0/21 0/21 Random combinatorial NQ10OVH NQ10 Vr B 1-8 VH B1-8 V% Nql0OVH Bl1-8 V% B 1-8 VH NQ1 V'K 2/49 11/103 0/44 10/24 27/49 56/103 40/44 10/49 10/49 14/103 22/103 2/44 2/44 4/24 4/24 6/24 PCR ,emplates were derived from a 1:1 mix of NQ1O/12.5 and B1-8 cells which were dividt-d into two portions. One portion of the cells were fixed before cDNA synthesis and PCR assembly and nested amplification (Fixed cells), and the other was boiled in water for cDNA synthesis and PCR assembly and nested amplification from soluble mRNA (Random combinatorial). Colonies were screened by PCR for Expts 1, 2 and 3.

but extra Expt. 1 colonies were also probed.

WO 93/03151 -4 7- PCT/GB92/01483 Table 3. Screening of linked VH and VL gene combinations from 1:9 mixtures of B1-8: NQ10/12.5 hybridoma cells.

PCR template Clones with Clones with B 1-8 VH (probed) B1-8 VH and B1.8

V?

Clones with B 1-8 VH and NQ10/12.5

VK

Fixed cells Random combinatorial 27/450 34/450 27/450 3/450 0/450 31/450 For preparation of PCR assembly templates a 1 9 mix of B 1-8 NQ10/12.5 cells was divided into two portions. One portion was fixed for cDNA synthesis and PCR (Fixed cells), and the other was boiled in water for cDNA synthesis and PCR from soluble mRNA (Random combinatorial). Clones were probed with 32 P labelled V186.2 probe, and those clones identified with B 1.8 VH genes were PCR screened for either V% or VK genes.

i i WO 93/03151 PCT/G B92/01 483

REFERENCES

1. Wick, M R, Loy, T, Mills, S E, Legier, J F, and Manivel, J L (1990). Hum.Path., 21, 758-766.

2. Waldmann, T A, (1991). Science, 252, 1657-1662.

3. Hamnxarstrom, S, Shively, J E, Paxton, R J, Beatty, B G, Larsson, A, Ghosh, R, et al. (1989). Cancer Res., 49, 4852-4858.

4. Barber, K E, Grosier, P S, Purdie, K J, Buchanan, J M, Cattermole, J A, Watson, J D, and Gillis, S (1989).

Growth Factors, 1, 101-114.

Orlandi, R, Gissow, D H, Jones, P T, and Winter, G (1989). Proc. Natl. Acad. Sci. USA, 86, 3833-3837.

6. Sastry, L, Alting-Mees, M, Huse, D W, Short, J M, Hay, B N, Janda, K D, Benkovic, S J, and Lerner, R A (1989). Proc. Natl. Acad. Sci. USA, 86, 5728-5732.

7. Ward, E 5, Gussow, D, Griffiths, A D, Jones, P T, and Winter, G (1991). Nature, 349, 293-299.

8. Huse, D W, Sastry, L, Iverson, S A, Kang, A S, Alting-Mees, M, Burton, D R, Benkovic, S J, and Lerner, R A (1989). Science, 246, 1275-1281.

9. McCafferty, J, Griffiths, A D, Winter, G, and Chiswell, D J (1990). Nature, 346, 552-554.

Clackson, T, Hoogenboom, H R, Griffiths, A D, and Winter, G (1991). Nature, 352, 624-628.

WO 93/03151 PCT/G B92/0 1483 -49- 11.. Winter, G, and Milstein, C, (1991). Nature, 349, 293-299.

12. Gherardi, E, and Milstein, C, (1992). Nature, 357, 201-202.

13. Kang, A S, Barbas, C F, Janda, K D, FBenkovic, S J, and Lerner, R A (1991). Proc. Natl. Acad. Sci. USA, 88, 4363-4366.

14. Zebedee, S L, Barbas, C F III, Horn, Y L, Caothien, R H, Graff, R, DeGraw, J, Pyati, J, LaPolla, R, Burton, D0R Lerner, R A and Thornton, G B (1992). Proc. Natl. Acad.

Sci. USA, 89, 3175-3179.

Haase, A T, Retzel, E F, and Staskus, K A (1990).

Proc. Natl. Acad. Sci. USA, 87, 4971-4975.

16. Cumano, A and Rajewsky, K (1985). Eur. J Immunol., 512-520. Nature, 312, 272-276.

17. Griffiths, C M, Berek, C, Kaartinen, M, and Milstein, C (1987). Nature, 312, 272-276.

18. Andersson, L C, Nilsson, K, and Gahmberg, C G (1979).

Int. J. Cancer, 23, 143-147.

19. Heisterkamp, N, Stem, K, Groffen, J, de Klein, A, and Grosveld, G (1985). Nature, 315, 758-761.

Ausubel, F M, Brent, R, Kingston, R E, Moore, D D, Seidman, D D, Smith, J A, and Struhi, K (1990). Current Protocols in Molecular Biology, Greene Publishing WO093/03151 PCT/GB92/01483 Associates and Wiley-Interscience, New York.

21. Thompson, J D, Brodsky, I, and Yunis, J J (1992).

Blood, 79, 1629-1635.

22. Davis, G T, Bedzyk, W D, Voss, E W, and Jacobs, T W (1991). Biotechnology, 9, 165-169.

23. Albert, J and Fenyo, E M (1990). J. Clin.

Microbiol., 28, 1560-1564.

24. Marchuk. D, Drummn, M, Saulino, A, and Collins, F S (1991). Nuci. Acids Res., 19, 1154.

Miller, J H (1972) Experiments in Molecular Genetics, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York.

26. Horowitz, J P, Chua, J, Curby, R J, Tomson, A J, Da Rooge, M F Fisher, B E, Mauricio, J, and Klundt, 1 (1964).

J. Med. Chem., 7, 574-575.

27. Cho, S, Scharpf, S, Franko, M, and Vermeulen, C W (1985). Biochem. Biophys. Res. Comm., 128, 1268-1273.

28. Gussow, D and Clackson, T (1989). Nuci. Acids Res., 17, 4000.

29. Bulwela, L, Forster, A, Boehm, T, and Rabbitts, T H (1989). Nuci. Acids Res., 17, 452.

Chebab, F F, and Kan, Y W (1989). Proc. Natl. Acad.

Sci. USA, 86, 9178-9182.

RA4

LU

0'4 Sf SUBSTITUTE SHEET :l I I: r* WO 93/03151 PC/GB92/01483 51 31. Bagasra, 0, Hauptman, S P, Lischner, H W, Sachs, M, and Pomerantz, R J (1992). New Engl. J. Med., 326, 1385- 1391.

32. Li, H H, Gyllensten, U B, Cui, X F, Saiki, R K, Erlich, H A and Arnheim, N (1988). Nature, 335, 414-417.

33. Skerra, A, Pfitzinger, I, and Pluckthun, A (1991).

Biotechnology, 9, 273-278.

34. Better, M, Chang, C P, Robinson, R R, and Horwitz, A H (1988). Science, 240, 1041-1043.

Novotny, J, Ganju, R K, Smiley, S T, Hussey, R E, Luther, M A, Recny, M A, Siliciano, R F, and Reinherz, E L (1991). Proc. Natl. Acad. Sci., USA, 88, 8646-8650.

36. Soo Hoo, W F, Lacy, M J, Denzin, L K, Voss, E W, Hardman, K D, and Kranz, D M (1992). Proc. Natl. Acad.

Sci. USA, 89, 4759-4763.

37. Ward, E S (1992). J. Mol. Biol., 224, 581-579.

38. Marks, J D, Hoogenboom, H R, Bonnert, T P, McCafferty, J, Griffiths, A D and Winter, G (1991). J.

Mol. Biol., 222, 581-579.

39. Li et al. (1988). Nature 345, 414-417.

Engelke et al. (1990). Anal. Biochem. 191, 396-400.

41. Hoogenboom, H R, Griffiths, A D, Johnson, K S, Chiswell, D J, Hudson, P and Winter, G (1991). Nucl.

Acids Res., 19, 143-151.

-f" U~:h A (Av o

Claims (15)

1. A method of treating a heterogeneous population of cells, the method comprising the step of linking together copies of two or more non-contiguous DNA sequences from at least some of the cells, the arrangement being such that copies of the DNA sequences from an individual cell are preferentially linked in the vicinity of the nucleic acid from which the copies are derived.

2. A method of treating a population of cells to link together copies of two or more non-contiguous DNA sequences from at least some of the cells, comprising r treating the cells to stabilise them with respect to temperature and to permeabilise them with respect to 15 reagents; adding primers for performing the polymerase chain reaction (PCR) such that the primers diffuse into the interior of the cells; subjecting the cells to suitable treatment so the DNA sequences of a particular cell are copied within that cell by PCR; and linking together the copies of the DNA sequences in the vicinity of the nucleic acid from which the copies are derived.

3. A method according to claim 1 or 2, wherein the 4 DNA sequence copies are linked by PCR. 1

4. A method according to claim 2 or 3, wherein the I1 DNA sequence subjected to PCR is derived from cellular mRNA. A method according to claim 4, wherein the DNA sequence is cDNA derived by reverse transcription of i mRNA.

6. A method according to any one of the preceding claims, further comprising a washing step following linking of DNA.

7. A method according to ly one of the preceding claims, further comprising one or more rounds or further rounds of PCR.

8. A method according to claim 7, wherein further rounds of PCR are performed using "nested" primers. S:20549c 53

9. A method according to any one of the preceding claims, wherein the DNA sequences comprise sequences encoding immunoglobulin VH and VL domains. A method according to any one of the preceding claims, wherein the resulting DNA is cloned by insertion into a suitable vector.

11. A method according to claim 10, whereby the clones are sequenced or annealed to nucleic acid probes.

12. A method according to claim 10 or 11, further comprising expressing a protein encoded by the DNA of a clone.

13. A method according to claim 12 wherein the protein is expressed as part of a fusion protein. till I 14. A method according to claim 13, wherein the 15 fusion protein comprises the coat protein of a filamentous phage or a peptide tag.

15. A protein expressed by the method of any one of F claims 12, 13 or 14. F 16. A composition comprising copies of two or more *CCU i 20 non-contiguous DNA sequences preferentially linked in the vicinity of the nucleic acid sequences from which the copies were derived, the preferentially linked copies being made by the method of any one of claims 1 to 11.

17. A method of treating a heterogeneous population of cells, the method comprising the step of linking together copies of two or more non-contiguous DNA sequences from at least some of the cells, the arrangement being such that copies of the DNA sequences i from an individual cell are preferentially linked in the vicinity of the nucleic acid from which the copies are derived, substantially as herein described with reference to the accompanying Examples.

18. A method of treating a population of cells to link together copies of two or more non-contiguous DNA sequences from at least some of the cells, comprising treating the cells to stabilise them with respect to temperature and to perneabilise them with respect to reagents; adding primers for performing the polymerase S:20549c 54 chain reaction (PCR) such that the primers diffuse into the interior of the cells; subjecting the cells to suitable treatment so the DNA sequences of a particular cell are copied within that cell by PCR; and linking together the copies of the DNA sequences in the vicinity of the nucleic acid from which the copies are derived, substantially as herein described with reference to the i accompanying Examples.

19. A composition comprising copies of two or more non-contiguous DNA sequences preferentially linked in the vicinity of the nucleic acid sequences from which the copies were derived, the preferentially linked copies I being made by the method of any one of claims 1 to 11, I t ,said composition being substantially as herein described 15 with reference to the accompanying examples. ,DATED this llth day of October 1995 I MEDICAL RESEARCH COUNCIL 1& By their Patent Attorneys GRIFFITH HACK CO. €c t Air0 I I l U 4/7-0 S:20"49c